Proc. Nati. Acad. Sci. USA Vol. 87, pp. 6502-6506, September 1990 Biochemistry Activation of murine c- protooncogene: Effect of a point on oncogenic activation (Abelson leukemia virus/Hardy-Zuckerman feline sarcoma virus/ kinase) SCOTT K. SHORE*, STEPHANIE L. BOGART, AND E. PREMKUMAR REDDY The Wistar Institute of Anatomy and Biology, 3601 Spruce Street, Philadelphia, PA 19104-4268 Communicated by Sidney Weinhouse, June 4, 1990 (receivedfor review February 15, 1990)

ABSTRACT Activation of the c-abl protooncogene occurs codons derived from the gag (4, 5). In addition to the 5' in Abelson murine leukemia virus, in Hardy-Zuckerman 2 , the P120 strain has undergone an internal in-frame feline sarcoma virus, and during the chromosomal transloca- deletion of 789 base pairs (bp) (263 codons) and a single point tions that generate BCR-ABL gene fusion products. To study mutation (G -* A) immediately downstream of the deletion the molecular mechanism involved in the c-abl activation, we [codon 832 of c-abl (type I) cDNA used in these studies] (4). have created a series of modifications in murine c-abl and Thus three structural changes, two deletions and a point assayed these constructs for oncogenic activity using the NIH mutation, appear to have accompanied the formation of the 3T3 cell transformation assay. Our results show that amino- P120 genome. The P160 genome has a very similar structure terminal deletions are sufficient for oncogenic activation of to that ofP120 but lacks the internal deletion seen in P120 (6). c-abl and high levels of oncogenic activities were generated by Earlier sequence studies had indicated the presence of a a deletion of 114 codons from the 5' end that deleted the SH3 frame-shift mutation in the P160 genome in the region that is region. A deletion of 53 codons from the 5' end (inclusive of deleted in the P120 genome (5). To confirm this observation, deletions seen in Hardy-Zuckerman 2 feline sarcoma virus and we resequenced the P160 DNA clone used in these studies, BCR-ABL gene products) that retains the SH3 region of c-abl and our sequence data (unpublished data) indicated that this resulted in the generation oflow levels oftransforming activity. frame-shift mutation is not present in the P160 clone that is This transforming potential was substantially increased with used here, suggesting that this frame-shift mutation seen is the introduction of a G -- A point mutation in codon 832 that either a sequencing error or a mutation that occurs in certain is present in v-abl. The point mutation was found to affect the DNA clones of P160 but does not contribute to the trans- secondary structure and the tyrosine kinase activity of the forming potential ofthe P160 genome. Thus, the P160 genome mutant gene products. has undergone two alterations during its oncogenic activa- tion, one of which is a deletion at the 5' end and the other a The v-abl gene was first identified as the transforming point mutation (at position 832 of the corresponding c-abl component of the Abelson murine leukemia virus (Ab- cDNA clone) of the encoded protein. MuLV), a murine retrovirus that causes B-cell leukemia in The HZ2-FeSV genome has been found to have suffered mice. Several different isolates of Ab-MuLV have been both 5' and 3' deletions, where a stretch of51 codons derived identified, and two ofthem, termed the P120 and P160 strains, from the 5' end has been replaced by a stretch of 344 codons have been extensively studied during the past few years. of the viral gag gene and a stretch of633 codons at the 3' end More recently, a second retrovirus, termed Hardy- has been replaced by 200 codons derived from the pol gene Zuckerman 2 feline sarcoma virus (HZ2-FeSV), was also of the helper virus (7). As the feline c-abl gene sequence has found to have transduced the feline abl gene in the generation not been determined, it is at present unclear as to whether the ofan acute transforming virus. In addition to its activation by feline v-abl gene has undergone any point in viral transduction, the ABL gene is also activated in Phila- addition to these gross deletions. delphia -positive (Phi') human chronic myelog- The activation ofthe humanABL gene in Phi' CML occurs enous leukemias (CMLs) and some forms of Phi' acute as a result of chromosomal translocation between chromo- lymphocytic leukemias (ALLs), which are characterized by somes 9 and 22 resulting in the replacement of 26 codons 9 and 22. derived from the 5' end of the ABL gene by 927 codons a reciprocal translocation between derived from the BCR gene (8, 9). Sequence analysis of a This translocation results in the fusion of two cellular , cDNA clone derived from the K562 cell line indicates there the ABL gene and the BCR gene, resulting in the formation ofthis of a chimeric gene termed BCR-ABL (reviewed in refs. 1 and are no additional mutations in the c-abl coding portion 2). gene (10). Thus, in all the three systems, the activation of the A comparative study of retroviral oncogenes and their c-abl gene was found to be associated with the deletion of a cellular homologues has provided important information re- varying proportion of 5' coding sequences derived from garding the structural changes that protooncogenes undergo exons 1 and 2 of the c-abl gene and their replacement with during the acquisition of transforming potential. Toward this either viral or BCR sequences. and An important biological consequence of these deletions goal, we had earlier reported the cloning sequence and mutations is seen in their relative transforming potential. analysis of Ab-MuLV proviral genome (P120 strain) as well have as the murine c-abl cDNA (3, 4). A comparison of the While the Ab-MuLV and HZ2-FeSV gene products nucleotide sequence of c-abl with its activated counterparts, been known to be potent transforming agents (7, 11), the P160 and P120, showed that both forms ofthe v-abl gene have BCR-ABL gene product is weakly transforming (12), being suffered a loss of their 5' sequences where 114 codons derived from the 5' end of c-abl have been replaced by 240 Abbreviations: Ab-MuLV, Abelson murine leukemia. virus; Mo- MuLV, Moloney murine leukemia virus; HZ2-FeSV, Hardy- Zuckerman 2 feline sarcoma virus; Phi', - The publication costs ofthis article were defrayed in part by page charge positive; ALL, acute lymphocytic leukemia; CML, chronic mye- payment. This article must therefore be hereby marked "advertisement" logenous leukemia. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 6502 Downloaded by guest on September 30, 2021 Biochemistry: Shore et al. Proc. Natl. Acad. Sci. USA 87 (1990) 6503 unable to transform NIH 3T3 cells in vitro (13). This observed initiation sites are unaltered. When it was necessary to difference in transforming activity could be due to one or all introduce a drug-resistance marker, the coding sequences ofthe three structural differences observed between the three were introduced into a murine retroviral vector, pMV7, activated forms ofabl. (i) The amino-terminal end ofthe viral which carries a selectable G418 drug-resistance marker (15). gag gene contains the site for myristoylation that allows This vector, which is derived from Mo-MuLV and has a efficient localization of the transforming protein to the mem- similar structure as that ofAb-MuLV, was previously shown brane. (ii) Deletions and point mutation seen in the v-abl gene to express v-abl and c-abl gene products efficiently and has that are not seen in the BCR-ABL and c-abl gene products been used by us and others in transformation assays (12, 13). contribute to the oncogenic potential of the v-abl gene Usage of these two vectors allowed us to directly test the product. (iii) The extent of amino-terminal truncation (26 vs. transforming potential of the cloned viral DNAs in an NIH 51 vs. 114 amino acids) seen between the three activated 3T3 assay and to rescue the transforming virus to test its forms contributes to the varying degree of transforming ability to transform cells in vitro and in vivo. The presence of activity seen with the three gene products. To discern be- the drug-resistance marker in the pMV7 vector also allowed tween these alternatives, we systematically attempted to us to monitor the relative efficiency obtained with each create deletions and mutations in the c-abl coding sequences transfection assay independent of focus formation, which and assessed their effect on the transforming activity of the was essential in the case of nontransforming constructs. The resulting chimeric genes. Our results suggest that activation presence of Xmn I and Bsm I restriction enzyme cleavage occurs with a deletion of 5' abl sequences and their replace- sites at the 5' end of c-abl cDNA clone provided us with a ment with gag sequences, whereas the point mutation ob- convenient way to create 5' deletions and chimeric hybrids served in the v-abl gene enhances the transforming potential between v-abl and c-abl DNA clones. The structure of the of c-abl with an altered 5' end. various mutant forms of the abl gene studied in this commu- nication is presented in Fig. 1. MATERIALS AND METHODS Transforming Potential of c-abl and gag-abl Fusion Prod- Cells and DNA Transfection Assays. NIH 3T3 cells were ucts. The transforming potential of the various mutant forms grown in Dulbecco's modified Eagle's medium containing Kbp 10% calf serum (GIBCO) and antibiotics. One hundred nano- grams of linearized plasmid DNA was transfected as de- 0 1 2 3 4 5 scribed (14). Individual foci were cloned by glass ring SH SH2542 TK trypsinization and expanded for analysis, or G418-resistant Focih/pa colonies were mass cultured. ATG X B Xs" E Sa TAG Bam c - a b abl Plasmid Construction. c-abl type I (4), Ab-MuLV P120 i- --:-:--- (3), Ab-MuLV P160 (6), and the retroviral vector pMV7 (15) ATGX B aX6 E Sa' TAG Barn I II were used for the construction of various mutants described. I . .. A6 1 -c-ab The complete c-abl (type I) reading frame was inserted into ATG X B Xho E Sal TAG Bam lI the vector pMV7 by digestion with EcoRI and BamHI gag-c-abi followed by blunting with the Klenow fragment of DNA polymerase I. A vector derived from Ab-MuLV proviral ATG X B K Sa. TAG Bar b genome (pC-1) was constructed by digestion with BstEII and I-K~xi,x cv-a b I-25 263-K HindIII, blunt-ending, and addition ofEcoRI linkers. For the ATG X B Xho E Sal TAG Bam construction ofgag-c-abl construct, this vector was digested - with EcoRI and BamHI, and the EcoRI to BamHI fragment -1 i gag--%S3-c-ab 1 5 of c-abl (I) was inserted. This allowed in-frame fusion of the Moloney murine leukemia virus (Mo-MuLV) p16 gag gene ATG X B X-o K Sal TAG Barn with the c-abl coding sequences. The same gene coding for NIXXXXXXNIX T Wvd gag-A53-c-abl-K gag-c-abl was inserted into pMV7 following ATG X B K Sal TAG Ba g BB digestion with Pvu I and BamHI, blunt-ending, and ligation 9 5 into the EcoRI site of the vector. Clone A61-c-abl has a 183-base 5' deletion of c-abl and was constructed by the ATG B Xho B Sal TAG Bam 1. I-t gag-5114-c-abl 370 addition of an ATG-containing linker to a prematurely ter- I minated cDNA clone and insertion of the resulting open ATG B K Sal TAG Ba- 5 1 0 reading frame into pMV7. 11111119 He P 1 20 Ab-MuLV-based constructs were made using the viral promoter elements and the p15 gag ATG initiation codon P 1 2 0 - E 4 7 0 present in the pC-1 vector. Chimeras between v-abl and c-abl clones were constructed by reciprocal exchange offragments ATG B XS-c K Sai TAG Ba'n generated by digestion of the cloned DNAs with Bsm I and P1 60 5 4 0 BamHI. Immunoprecipitation and Kinase Assay. Immunoprecipita- FIG. 1. Construction of retroviral vectors expressing various tion and kinase assays were performed as described by forms of abl gene.WJ Details~~~~~~~~~~~~~~gag--%53-263-c-ab1-Kof vector construction are provided in the text. The SH3, SH2, and kinase regions are indicated. Clones ofc-abl Konopka and Witte (16). The peptide-derived anti-v-abl cDNA (type I), gag-c-abl, A61-c-abl, and c-abl-A263-K were inserted antiserum was kindly provided by R. Arlinghaus and has into the retroviral vector pMV7, whereas the remaining constructs been described (17). were made in Ab-MuLV. Dark boxes represent viral gag sequences, hatched boxes represent v-abl sequences, and gray boxes represent RESULTS the c-abl sequences. B, Bsm I; X, Xmn I; Xho, Xho I; Sal, Sal I; Bam, BamHI. The location of the mutation is shown by the normal Construction of Retroviral Vectors Coding for c-abl, v-abl, glutamic acid residue (E) or the mutant lysine amino acid (K), and the and Mutant Forms of c-abl. Construction of recombinants deletions are indicated by bent lines. One hundred nanograms of between v-abl and c-abl genes was carried out with the linearized proviral DNA was transfected onto NIH 3T3 cells, and the Ab-MuLV proviral DNA clone, such that the viral long results have been multiplied by 10 and expressed as focus formation terminal repeat structure and transcription and translation per Ag of DNA. Kbp, kilobase pairs. Downloaded by guest on September 30, 2021 6504 Biochemistry: Shore et A Proc. Natl. Acad. Sci. USA 87 (1990) of the abl gene was tested by using the NIH 3T3 transfor- mation assay, as described (14). In all instances, the trans- -. I i A - formation assays were carried out in the absence of any I .t: Mo-MuLV DNA so that the true transformation efficiency could be scored without the formation of any secondary foci 7Of due to the spread of the virus and to prevent recombination between the two DNAs. The transforming activities obtained with various constructs are presented in Fig. 1. The results show that a retroviral vector expressing c-abl alone was incapable of transforming NIH 3T3 cells. Similarly, deletion of 61 codons from the 5' end (a stretch of deletion inclusive of the deletions seen in the HZ2-FeSV and BCR-ABL gene products) by itself had no effect on the transforming potential of c-abl as these DNA constructs also showed no focus formation in the NIH 3T3 assay. In addition, fusion of the complete c-abl coding sequences to gag sequences of Mo- MuLV also had no enhancing effect on the transforming potential of c-abl. Introduction of the internal deletion and point mutation present in the P120 genome without altering the 5' end also had no activating effect on the c-abl genome. We next tested the effect of 5' deletions and their replace- ment with the viral gag sequences on the transforming poten- tial ofthe c-abl gene. Thus a deletion of 53 codons from the 5' end (comparable to the deletion seen in HZ2-FeSV and inclusive ofthe deletion seen in BCR-ABL gene products) and their replacement with 35 viral gag codons resulted in the generation of low levels of transforming activity of the gag- A&53-c-ablconstructs. However, by eliminating 114 codons and replacing them with gag sequences (as is seen with two Ab-MuLV proviral DNAs and gag-A114-c-abb, this trans- forming potential substantially increased to a level seen with the proviral Ab-MuLV DNAs (Fig. 1). We next examined whether the point mutation and deletion seen in Ab-MuLV proviral genome had any synergic effect on the transforming potential of the various deletion mutants studied above. Our results, presented in Fig. 1, show that the simple introduction of the G -+ A point mutation in gag-A53- c-abl (where 53 codons at the 5' end have been replaced by the viral gag sequences) resulted in a dramatic increase in its transforming potential as is seen with gag-A53-c-abl-K. In- troduction of the point mutation and the internal deletion of codons 542-805 seen in the Ab-MuLV P120 genome (gag- A53-263-c-abl-K) resulted in a slight but not significantly higher increase in the transforming potential compared to the introduction of point mutation alone. Interestingly, the effect of the point mutation on NIH 3T3 transformation is not apparent with the 114-codon deletion ofAb-MuLV, as shown by the transformation efficiency ofgag-A114-c-abl DNA. We confirmed this by creating the site-specific back mutation (A -* G) in Ab-MuLV P120 and finding no difference in trans- formation efficiency compared with wild-type P120 (Fig. 1). Fig. 2 shows the morphology of cell lines that were isolated from foci or G418-resistant colonies. NIH 3T3 cells trans- fected with c-abl constructs or its nontransforming deletion mutants showed a flat morphology and closely resembled the parental NIH 3T3 cells. However, the transforming mutants described in Fig. 1 generated cell lines that were highly refractile and spherical, as is seen with the two Ab-MuLV proviral DNAs. The only exception was the cells transformed by gag-A53-c-abl DNA, which were more fusiform and less refractile than the Ab-MuLV-transformed cell lines. Intro- FIG. 2. Morphology of NIH 3T3 cells isolated following trans- duction of a point mutation in this DNA not only resulted in fection with various abl constructs. (A) NIH 3T3. (B) A61-c-abl. (C) increased transforming potential of this DNA clone but also gag-A&53-c-abl. (D) gag-A53-263-c-abl-K. (E) gag-A53-c-abl-K. (F) in the alteration of the morphology of the transformed cells gag-A114-c-abl. (G) Ab-MuLV P120. (H) Ab-MuLV P160. (Photo- into a more malignant phenotype. graphed at x200 with Hoffman modulation contrast.) Analysis of the Tyrosine Kinme Activity of the Chimeric abl . The abl gene products have earlier been shown to tion potential of various mutants of abl correlated with their possess an associated tyrosine kinase activity that is substan- kinase activity. It was especially interesting to see if the point tially enhanced following its oncogenic activation (18-20). We, mutation had a synergistic effect that correlated with its therefore, analyzed whether the variability in the transforma- increased transforming activity. The results presented in Fig. Downloaded by guest on September 30, 2021 Biochemistry: Shore et al. Proc. Natl. Acad. Sci. USA 87 (1990) 6505

3 show that the cell lines transfected with the nontransforming Mr X 10-3 1 2 3 4 5 6 7 constructs (lanes 2-5) failed to show any kinase activity in immune complex kinase assays (16). On the other hand, cell 77 - lines transfected with highly transforming mutants of abl, P120, P160, and gag-A114-c-abl constructs showed potent autophosphorylation as well as phosphorylation of cellular proteins (lanes 9-11). The constructs with lesser transforming activity, gag-A53-c-abl-K and gag-A53-263-c-abl-K (lanes 7 and 8), showed autokinase activity similar to the more potent transforming species. However, it is interesting to note that 46.5 _ with these mutants the phosphorylation of non-abl proteins is Enolase ' considerably less compared to that seen with mutants with more potent transforming activity. Cells transfected with gag-A53-c-abl constructs expressed a protein with extremely low kinase activity (Fig. 3, lane 6), which could only be seen with prolonged exposure (data not FIG. 4. Phosphorylation of acid-denatured enolase (20) by cell shown). However, the introduction of the point mutation at lysates from NIH 3T3 cells transfected with various DNA constructs. Cell lysates were equilibrated for protein concentration and immu- codon 832 of c-abl in this clone had a profound enhancing noprecipitated with anti-v-abl TL037 antiserum. To the immune effect on its kinase activity (Fig. 3, lane 7), suggesting that complex 10 ,Ci of [y.32P]ATP and 1 p1 of rabbit muscle acid- enhanced transformed activity seen with this mutant corre- denatured enolase were added. The reaction mixture was incubated lated with its increased kinase activity. for 15 min at room temperature and analyzed on a 6% SDS/PAGE To further confirm this observation, we next determined the gel. Autoradiography of the dried gel was for 20 min. Lane 1, NIH ability of various mutants to phosphorylate an exogenous 3T3 cells; lane 2, A61-c-abl; lane 3, gag-A53-c-abl; lane 4, gag-A53- substrate, such as enolase (21). The results of this study c-abl-K; and lane 5, gag-A53-263-c-abl-K. Lanes 6 and 7 are the same presented in Fig. 4 demonstrate that neither NIH 3T3 cells as lanes 3 and 4, but the exposure time was 40 min. (lane 1) nor the nontransforming mutant of c-abl (lane 2) was mutation seen in Ab-MuLV proviral DNA acts to increase the able to phosphorylate enolase. A very low level of enolase transforming potency by enhancing the kinase activity. phosphorylation was observed with cell extracts transformed by gag-A53-c-abl (lane 3), which is also shown with prolonged exposure of the autoradiograms (lane 6). However, with the DISCUSSION introduction of the point mutation or point mutation and Results presented in this communication demonstrate that deletion (lanes 4, 5, and 7) into these DNA clones, a dramatic the oncogenic activation ofthe c-abl gene is influenced by the increase in their kinase activity was observed. The levels of extent of amino-terminal deletions as well as by an internal enolase phosphorylation observed closely correspond to the point mutation. Thus, full oncogenic potential of this onco- transforming activity exhibited by these various mutant gene, as measured by NIH 3T3 transformation assay, can be DNAs, further supporting the observation that the point achieved by a deletion of 114 codons from the 5' end of the 1 2 3 4 5 6 7 8 9 10 11 c-abl gene and their replacement by viral gag sequences. The 5' end of the c-abl gene (see Fig. 1) has previously been shown to contain two regions of homology with the c-src Mr X 10-3 sequences, which have been designated as SH3 and SH2

* ...* *..& regions. Not only are these sequences present in all src- related kinases but also the SH3 domain has been found in 130 - phospholipase C, v-crk, GAP, a-spectrin, myosin I, and a yeast actin binding protein (reviewed in ref. 22; ref. 23). The * _ SH3 region is thought to play the role ofa negative regulatory 75 - element since the oncogenic activation of c-src has been .' observed following the introduction ofpoint mutations in the *4_ SH3 region (24, 25). The presence of viral gag sequences in the mutants studied here seems to provide an important structural requirement-namely, the provision of a myris- 50- toylation signal that allows efficient membrane localization of _.. ;s the oncogenic product. This localization of the abl gene product has recently been demonstrated to be characteristic of the transforming protein (26). Our results also demonstrate that when a limited set of FIG. 3. Immune complex kinase assay of abl proteins in cell sequences is deleted from the 5' end of this gene, there is a lysates from transfected NIH 3T3 cells (16). abl proteins from cell considerable reduction in the transforming potential of the lysates containing equal amounts of protein were immunoprecipi- tated with a-v-abl TL037 antiserum and collected by the addition of mutant gene. Thus, replacement of the first 53 codons with protein A-Sepharose beads. Immune complexes were suspended in p15 viral gag sequences resulted in the generation of a low 15 ,ul of kinase buffer (20 mM Hepes, pH 7.0/0.1% Triton X-100/20 level of transforming activity. However, this transforming mM MnCl2/150 mM NaCl) and labeled for 15 min at room temper- activity was considerably enhanced by the introduction of a ature following the addition of 10 gCi (1 Ci = 37 GBq) of [y-32P]ATP G -+ A point mutation in codon 832 of the c-abl gene, a point in 3 ,ul ofbuffer. The proteins were resolved on a 6% SDS/PAGE gel mutation that results in the substitution oflysine for glutamic and exposed to film for 2 min. The autophosphorylated abl proteins acid, and was originally seen by us in the v-abl gene se- are indicated by the asterisks. Lane 1, NIH 3T3; lane 2, c-abl; lane to have a 3, gag-c-abl; lane 4, c-abl-A263-K; lane 5, A61-c-abl; lane 6, gag- quences (4). Thus, this point mutation appears A53-c-abl; lane 7, gag-A53-c-abl-K (Mr 140,000); lane 8, gag-A53- synergistic effect on the oncogenic activation of this gene, 263-c-abl-K (Mr 90,000); lane 9, P12 (Mr 120,000); lane 10, gag-A114- especially in instances where only a limited set of sequences c-abl (Mr 160,000); and lane 11, p160 (Mr 160,000). is deleted from the 5' end of the gene. Downloaded by guest on September 30, 2021 6506 Biochemistry: Shore et al. Proc. Natl. Acad. Sci. USA 87 (1990) Biochemical analysis ofthe normal and mutant proteins for codon 832 are likely to potentiate its transforming activity. It their associated kinase activity demonstrates that oncogenic is possible that such mutations constitute a secondary event, activation of the c-abl gene product is associated with con- at least in some cases ofCML and ALL, that leads to a more comitant activation of the tyrosine kinase activity of the malignant disease. A careful sequence analysis of the BCR- deletion mutants. The mutants with a deletion of 114 codons ABL gene product from a large population of Phi' CML and from the 5' end of the gene exhibited a kinase activity ALL patients in blast and chronic phases of the disease comparable to that of wild-type v-abl gene products. The should clarify the nature of molecular events that lead to the mutant in which the first 53 codons were replaced with viral oncogenic activation of this gene in human leukemias. p15 gag sequences produced a gene product that exhibited extremely low levels of kinase activity, but the introduction We thank S. Goff for the gift of Ab-MuLV strain P160, B. of the G -- A point mutation in codon 832 in this deletion Weinstein for the pMV7 vector, M. LaCava for technical assistance, mutant resulted in a dramatic elevation of its kinase activity. and J. Brugge for the critical reading of this manuscript. This work Thus, this mutation appears to enhance the oncogenic po- is supported by National Institutes of Health Grant CA-47937-01. tential of the gag-A53-c-abl construct by synergistically en- 1. Reddy, E. P. (1988) in Oncogene Handbook, eds. Reddy, hancing the tyrosine kinase activity of the encoded protein. E. P., Skalka, A. M. & Curran, T. (Elsevier, Amsterdam), pp. The computer analysis of the c-abl amino acid sequence 3-23. using the Chou-Fasman plot structure program for secondary 2. Rosenberg, N. & Witte, 0. N. (1988) Adv. Virus Res. 35, structure (27) of gene products with and without the point 39-81. mutation at codon 832 shows that this mutation could result 3. Reddy, E. P., Smith, M. J. & Srinivasan, A. (1983) Proc. Natl. in an alteration of the secondary structure, where the turn- Acad. Sci. USA 80, 3623-3627. 4. Oppi, C., Shore, S. K. & Reddy, E. P. (1987) Proc. Natl. Acad. helix-turn structure seen with the normal gene product is Sci. USA 84, 8200-8204. disrupted (data not shown). The synergistic effects seen 5. Lee, R., Paskind, M., Wang, J. Y. J. & Baltimore, D. (1985) in between the 5' deletions and internal point mutation in RNA Tumor Viruses, eds. Weiss, R., Teich, N., Varmus, H. & generating the transforming potential and tyrosine kinase Coffin, J. (Cold Spring Harbor Lab., Cold Spring Harbor, NY), activity of the protein products suggest a possible intramo- pp. 861-868. lecular interaction between these two domains-i.e., the SH3 6. Goff, S. P., Gilboa, E., Witte, 0. N. & Baltimore, D. (1980) domain and the domain encoded by the region that includes Cell 22, 777-785. codon 832. It is possible that these two domains interact with 7. Bergold, P. J., Blumenthal, J. A., D'Andrea, E., Snyder, each other in their native state bringing about an intramo- H. W., Lederman, L., Silverstone, A., Nguyen, H. & Besmer, lecular negative regulatory effect on the protein P. (1987) J. Virol. 61, 1193-1202. molecule, 8. Heistercamp, N., Stephenson, J. R., Groffen, J., Hansen, and this interaction is disrupted either by extensive deletions P. F., de Klein, A., Bartram, C. R. & Grosveld, G. (1983) in the 5' end or a smaller deletion at the 5' end and concom- Nature (London) 306, 239-242. itant point mutation in codon 832. This turn-helix-turn struc- 9. Shtivelman, E., Lifshitz, B., Gale, R. P. & Canaani, E. (1985) ture is also conserved in the human ABL protein, indicating Nature (London) 315, 550-554. its functional importance (10). 10. Fainstein, E., Einat, M., Gokkel, E., Marcell, C., Croce, While this work was in progress, two reports appeared C. M., Gale, R. P. & Canaani, E. (1989) Oncogene 4, 1477- where the effect of 5' deletions and their replacement by viral 1481. gag sequences or deletions in the naturally 11. Scher, C. D. & Siegler, R. (1975) Nature (London) 253, 729- myristoylated 731. c-abl (type IV) were studied (28, 29). Our results agree with 12. McLaughlin, J., Chianese, E. & Witte, 0. N. (1987) Proc. Natl. theirs with respect to lack of transformation by c-abl and Acad. Sci. USA 84, 6558-6562. myristoylated c-abl proteins. However, these reports suggest 13. Daley, G. Q., McLaughlin, J., Witte, 0. N. & Baltimore, D. that complete deletion of the SH3 region is essential for the (1987) Science 237, 532-535. oncogenic activation of the c-abl gene. Our results differ from 14. Shore, S. K. & Reddy, E. P. (1989) Oncogene 4, 1411-1413. these two reports in that we observe oncogenic activation of 15. Kirshmeier, P. T., Housey, G. M., Johnson, M. D., Perkins, the c-abl gene without deletion of the SH3 domain as is seen A. S. & Weinstein, I. B. (1988) DNA 7, 219-225. in the gag-A53-c-abl, gag-A&53-c-abl-K, and gag-A53-263-c- 16. Konopka, J. B. & Witte, 0. N. (1985) Mol. Cell. Biol. 5, abl-K constructs. Furthermore, these investigators have not 3116-3123. 17. Kloetzer, W., Kuzrock, R., Smith, L., Talpaz, M., Spiller, M., studied the effects of point mutation in codon 832, which Gutterman, J. & Arlinghaus, R. (1985) Virology 140, 230-238. seems to impart an enhanced transforming activity over the 18. Witte, 0. N., Dasgupta, A. & Baltimore, D. (1980) Nature weakly oncogenic gag-A53-c-abl gene product. (London) 283, 826-831. The results presented in this communication also have 19. Konopka, J. B., Watanabe, S. M. & Witte, 0. N. (1984) Cell important implications for the oncogenic activation of human 87, 1035-1042. ABL in Phi' CML and Phi' ALL patients. The BCR-ABL 20. Besmer, P., Hardy, W. D., Jr., Zuckerman, E. E., Bergold, P., gene has been shown to code for a protein in which the first Lederman, L. & Snyder, H. W., Jr. (1983) Nature (London) 26 amino acids of the c-abl protein have been replaced by a 303, 825-828. stretch of amino acids derived 21. Cooper, J. A., Esch, F. S., Taylor, S. S. & Hunter, T. (1984) from the BCR gene (8, 9). It is J. Biol. Chem. 259, 7835-7841. interesting to note that the onset ofPhiladelphia . Dawson, T. (1988) Oncogene 3, 491-495. does not immediately result in the onset of CML or ALL. 23. Drubin, D. G., Mulholland, J., Zhu, Z. & Butstein, D. (1990) Patients who are Phi' live fairly normal lives for prolonged Nature (London) 343, 288-290. periods. However, an alteration in the nature of the disease 24. Potts, W. M., Reynolds, A. B., Lansing, T. J. & Parsons, J. T. occurs that seems to result from a single clone evolving into (1988) Oncogene Res. 3, 343-355. a more malignant cell type. At this acute phase of the illness, 25. Kato, J. Y., Takeya, T., Grandori, C., Iba, H., Levy, J. B. & survival of the patient is limited to a few months (reviewed in Hanafusa, H. (1986) Mol. Cell. Biol. 6, 4155-4160. ref. 30). The molecular events that dictate this transition are 26. Van Etten, R. A., Jackson, P. & Baltimore, D. (1989) Cell 58, at present unclear. The results presented in 669-678. this communi- 27. Chou, P. Y. & Fasman, G. D. (1978) Adv. Enzymol. 47,45-148. cation indicate why the BCR-ABL gene product, which has a 28. Franz, W. M., Berger, P. & Wang, J. Y. J. (1989) EMBO J. 8, minimal deletion of amino-terminal amino acids derived from 137-147. the ABL gene, is weakly transforming or even nontransform- 29. Jackson, P. & Baltimore, D. (1989) EMBO J. 8, 449-456. ing in NIH 3T3 cells (12, 13). Additional accumulation of 30. Kurzrock, R., Gutterman, J. U. & Talpaz, M. (1988) N. Engl. mutations in functional domains such as that in the vicinity of J. Med. 319, 990-998. Downloaded by guest on September 30, 2021